public inbox for linux-fsdevel@vger.kernel.org
 help / color / mirror / Atom feed
From: Chi Zhiling <chizhiling@163.com>
To: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org
Cc: Namjae Jeon <linkinjeon@kernel.org>,
	Sungjong Seo <sj1557.seo@samsung.com>,
	Yuezhang Mo <yuezhang.mo@sony.com>,
	Chi Zhiling <chizhiling@kylinos.cn>
Subject: [PATCH v2 1/6] exfat: add block readahead in exfat_chain_cont_cluster
Date: Tue,  3 Mar 2026 11:14:04 +0800	[thread overview]
Message-ID: <20260303031409.129136-2-chizhiling@163.com> (raw)
In-Reply-To: <20260303031409.129136-1-chizhiling@163.com>

From: Chi Zhiling <chizhiling@kylinos.cn>

When a file cannot allocate contiguous clusters, exfat converts the file
from NO_FAT_CHAIN to FAT_CHAIN format. For large files, this conversion
process can take a significant amount of time.

Add simple readahead to read all the FAT blocks in advance, as these
blocks are consecutive, significantly improving the conversion performance.

Test in an empty exfat filesystem:
  dd if=/dev/zero of=/mnt/file bs=1M count=30k
  dd if=/dev/zero of=/mnt/file2 bs=1M count=1
  time cat /mnt/file2 >> /mnt/file

| cluster size | before patch | after patch |
| ------------ | ------------ | ----------- |
| 512          | 47.667s      | 4.316s      |
| 4k           | 6.436s       | 0.541s      |
| 32k          | 0.758s       | 0.071s      |
| 256k         | 0.117s       | 0.011s      |

Signed-off-by: Chi Zhiling <chizhiling@kylinos.cn>
---
 fs/exfat/exfat_fs.h | 11 +++++++++--
 fs/exfat/fatent.c   | 37 +++++++++++++++++++++++++++++++++++++
 2 files changed, 46 insertions(+), 2 deletions(-)

diff --git a/fs/exfat/exfat_fs.h b/fs/exfat/exfat_fs.h
index 2dbed5f8ec26..090f25d1a418 100644
--- a/fs/exfat/exfat_fs.h
+++ b/fs/exfat/exfat_fs.h
@@ -10,6 +10,7 @@
 #include <linux/ratelimit.h>
 #include <linux/nls.h>
 #include <linux/blkdev.h>
+#include <linux/backing-dev.h>
 #include <uapi/linux/exfat.h>
 
 #define EXFAT_ROOT_INO		1
@@ -79,6 +80,10 @@ enum {
 #define EXFAT_HINT_NONE		-1
 #define EXFAT_MIN_SUBDIR	2
 
+#define EXFAT_BLK_RA_SIZE(sb)		\
+	(min_t(blkcnt_t, (sb)->s_bdi->ra_pages, (sb)->s_bdi->io_pages) \
+	<< (PAGE_SHIFT - (sb)->s_blocksize_bits))
+
 /*
  * helpers for cluster size to byte conversion.
  */
@@ -117,9 +122,9 @@ enum {
 #define FAT_ENT_SIZE (4)
 #define FAT_ENT_SIZE_BITS (2)
 #define FAT_ENT_OFFSET_SECTOR(sb, loc) (EXFAT_SB(sb)->FAT1_start_sector + \
-	(((u64)loc << FAT_ENT_SIZE_BITS) >> sb->s_blocksize_bits))
+	(((u64)(loc) << FAT_ENT_SIZE_BITS) >> sb->s_blocksize_bits))
 #define FAT_ENT_OFFSET_BYTE_IN_SECTOR(sb, loc)	\
-	((loc << FAT_ENT_SIZE_BITS) & (sb->s_blocksize - 1))
+	(((loc) << FAT_ENT_SIZE_BITS) & (sb->s_blocksize - 1))
 
 /*
  * helpers for bitmap.
@@ -448,6 +453,8 @@ int exfat_find_last_cluster(struct super_block *sb, struct exfat_chain *p_chain,
 		unsigned int *ret_clu);
 int exfat_count_num_clusters(struct super_block *sb,
 		struct exfat_chain *p_chain, unsigned int *ret_count);
+int exfat_blk_readahead(struct super_block *sb, sector_t sec,
+		sector_t *ra, blkcnt_t *ra_cnt, sector_t end);
 
 /* balloc.c */
 int exfat_load_bitmap(struct super_block *sb);
diff --git a/fs/exfat/fatent.c b/fs/exfat/fatent.c
index f87576ca7032..9a4143f3fc0c 100644
--- a/fs/exfat/fatent.c
+++ b/fs/exfat/fatent.c
@@ -142,13 +142,50 @@ int exfat_ent_get(struct super_block *sb, unsigned int loc,
 	return -EIO;
 }
 
+int exfat_blk_readahead(struct super_block *sb, sector_t sec,
+		sector_t *ra, blkcnt_t *ra_cnt, sector_t end)
+{
+	struct blk_plug plug;
+
+	if (sec < *ra)
+		return 0;
+
+	*ra += *ra_cnt;
+
+	/* No blocks left (or only the last block), skip readahead. */
+	if (*ra >= end)
+		return 0;
+
+	*ra_cnt = min(end - *ra + 1, EXFAT_BLK_RA_SIZE(sb));
+	if (*ra_cnt == 0) {
+		/* Move 'ra' to the end to disable readahead. */
+		*ra = end;
+		return 0;
+	}
+
+	blk_start_plug(&plug);
+	for (unsigned int i = 0; i < *ra_cnt; i++)
+		sb_breadahead(sb, *ra + i);
+	blk_finish_plug(&plug);
+	return 0;
+}
+
 int exfat_chain_cont_cluster(struct super_block *sb, unsigned int chain,
 		unsigned int len)
 {
+	sector_t sec, end, ra;
+	blkcnt_t ra_cnt = 0;
+
 	if (!len)
 		return 0;
 
+	ra = FAT_ENT_OFFSET_SECTOR(sb, chain);
+	end = FAT_ENT_OFFSET_SECTOR(sb, chain + len - 1);
+
 	while (len > 1) {
+		sec = FAT_ENT_OFFSET_SECTOR(sb, chain);
+		exfat_blk_readahead(sb, sec, &ra, &ra_cnt, end);
+
 		if (exfat_ent_set(sb, chain, chain + 1))
 			return -EIO;
 		chain++;
-- 
2.43.0


  reply	other threads:[~2026-03-03  3:15 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-03-03  3:14 [PATCH v2 0/6] exfat: improve performance of NO_FAT_CHAIN to FAT_CHAIN conversion Chi Zhiling
2026-03-03  3:14 ` Chi Zhiling [this message]
2026-03-03  3:14 ` [PATCH v2 2/6] exfat: use readahead helper in exfat_allocate_bitmap Chi Zhiling
2026-03-03  3:14 ` [PATCH v2 3/6] exfat: use readahead helper in exfat_get_dentry Chi Zhiling
2026-03-03  3:14 ` [PATCH v2 4/6] exfat: drop redundant sec parameter from exfat_mirror_bh Chi Zhiling
2026-03-03  3:14 ` [PATCH v2 5/6] exfat: optimize exfat_chain_cont_cluster with cached buffer heads Chi Zhiling
2026-03-03  3:14 ` [PATCH v2 6/6] exfat: fix error handling for FAT table operations Chi Zhiling
2026-03-05 12:15 ` [PATCH v2 0/6] exfat: improve performance of NO_FAT_CHAIN to FAT_CHAIN conversion Namjae Jeon

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260303031409.129136-2-chizhiling@163.com \
    --to=chizhiling@163.com \
    --cc=chizhiling@kylinos.cn \
    --cc=linkinjeon@kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=sj1557.seo@samsung.com \
    --cc=yuezhang.mo@sony.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox